'JavaPackage' object is not callable error in Python API

[sryza]

This was reported on the mailing list:
https://groups.google.com/forum/#!topic/spark-ts/4ZJ7t-w0s7o

<ipython-input-8-b40312bd4ec8> in <module>()
      1 # Create an daily DateTimeIndex over August and September 2015
----> 2 freq = BusinessDayFrequency(1, sc)

/databricks/python/local/lib/python2.7/site-packages/sparkts/datetimeindex.pyc in __init__(self, bdays, sc)
    115 
    116     def __init__(self, bdays, sc):
--> 117         self._jfreq = sc._jvm.com.cloudera.sparkts.BusinessDayFrequency(bdays)
    118 
    119     def __eq__(self, other):

[hajimupakura]

Did anyone get a solution to this error?

[blpabhishek]

if I am not wrong you are not providing a way for adding sparkts jar as PYSPARK_SUBMIT_ARGS to use jvm objects in python.

[diegorep]

This happened to me when I had not properly set the --jars path in the spark-defaults.conf.

My experience is that these "not callable" errors usually show up when an object does not have some property or when the property is being called with incorrect parameters. Here, the object is the com.cloudera.sparkts package and we'd technically be looking for property BusinessDayFrequency.

Usually, python will throw a different error if it can't find sparkts (most likely at some import statement), but since sparkts is a java package living in the JVM, it won't attempt this until you actually try to access BusinessDayFrequency, which will of course break it if the sparkts jar was not properly added to the spark conf.

@blpabhishek's solution will also work, albeit you'll have to make sure you pass in that argument each time you start up the shell